Salivary proteins and glycoproteins are the primary source of carbon and nitrogen for growth of oral bacteria in dental plaque biofilm communities. Little is known, however, about how different species degrade these macromolecules for use as growth substrates, and how the composition of biofilm communities is influenced by cross feeding between species. To address these questions, we initiated studies of bacterial growth and biofilm formation in filter-sterilized whole saliva. Our goal is to identify metabolic interactions between species that drive oral biofilm development. A standard protocol developed for growth experiments involves inoculation of sterile saliva with bacteria cultured in complex media and subsequent transfer of primary saliva cultures into fresh saliva to obtain secondary saliva cultures. Final growth of Streptococcus gordonii in primary and secondary saliva cultures was comparable and was associated with a 25 to 50% loss of salivary carbohydrate but little change in total salivary protein. In contrast, final growth of S. mutans was 30-times less in secondary than primary cultures, and was not associated with a measurable decrease in salivary carbohydrate content. Lectin-blotting of saliva from secondary cultures revealed degradation of salivary glycoproteins by S. gordonii but not by S. mutans. We are also exploring whether growth of S. gordonii promotes associated growth of cariogenic S. mutans, a possibility that has important implications for survival of the latter species in the oral environment. The genome of S. gordonii DL1 encodes three putative cell surface glycoside hydrolases (GHs), which we hypothesize are important for growth of this organism in saliva. We have now prepared a panel of wild type and mutant strains that lack (or express) one, two or all three cell-surface GHs. Growth studies with selected strains were performed in heat-treated saliva that was first dialyzed to remove possible low molecular weight growth substrates. Under such conditions, growth of the mutant lacking all three cell surface GHs was 10- to 20-fold less than the wild type. Smaller reductions in growth were noted with mutants lacking each individual GH. SDS-PAGE and lectin blotting of saliva from growth cultures revealed deglycosylation of specific salivary components, most notably one that migrated in the 70 kDa region, which was identified as a proline-rich glycoprotein. We are also comparing wildtype and mutant strains for degradation of O- and N-linked oligosaccharide chains on different model glycoproteins. In these experiments, S. gordonii DL1 deglycosylated asialofetuin but not fetuin, which indicates that terminal sialic acid protects oligosaccharide chains on the latter protein. From this, we wonder whether the presence of terminal sialic acid on salivary glycans limits growth of S. gordonii in saliva. If so, this streptococcal species, which does not produce sialidase, may benefit from its association with sialidase-producing organisms such as S. oralis or Actinomyces spp. in biofilm communities. In earlier studies, we noted that the arrangement of certain genes for RPS biosynthesis differed between strains of S. gordonii and S. oralis. Based on this, we wondered whether the arrangement of these genes can be used to distinguish different species of receptor polysaccharide (RPS)-producing streptococci. To address this question, we identified and compared loci for polysaccharide biosynthesis in additional strains of S. gordonii and S. oralis as well as S. sanguinis, a previously unstudied species. Such comparisons revealed clear differences between RPS-producing species. Thus, rml genes for synthesis of TDP-L-Rha were in rps loci of S. oralis strains but at other loci in S. gordonii and S. sanguinis. Moreover, galE1 for interconversion of UDP-Glc and UDP-Gal was in galactose operons of all S. gordonii and S. sanguinis strains but surprisingly, was not present in strains of S. oralis. Moreover, galE2 for interconversion of UDP-Glc and UDP-Gal as well as UDP-GlcNAc and UDP-GalNAc was found at a different locus in each species, including the rps locus in strains of S. sanguinis. These findings provide insight into metabolic properties that distinguish different RPS-producing streptococci and open a molecular approach for identifying these bacteria based on the arrangement of genes for synthesis of polysaccharide precursors. Importantly, a molecular approach to this problem complements those based on the use of RPS-specific antibodies, which do not distinguish different species. Results from previous studies showed that oral bacterial biofilms originate and develop as multi-species communities composed of organisms that bear complementary cell surface adhesins and receptors, such as type 2 fimbriated Actinomyces spp. and RPS-bearing streptococci. Results of these studies were limited however by the availability of antibodies against the different RPS serotypes produced by strains of S. gordonii, S. oralis and S. sanguinis. In current studies, antibodies against all five RPS serotypes are being used to verify and extend previous observations across a range of individuals. The presence of more than one RPS serotype on different cells within the same biofilm community has been noted. In addition, bacteria labeled with antibody against a unique rhamnose-containing type of RPS produced by S. oralis H1 have been identified in the plaque biofilm of one subject. Studies are underway to identify bacteria in the biofilm that bear complementary cell surface adhesins. Importantly, plaque biofilms formed in all subjects contain many bacteria that are not labeled by the available panel of specific antibodies. DNA extracted from plaque biofilms is being analyzed by Human Oral Microbiome Database microarray to provide a starting point for identification of unlabeled members of plaque biofilm communities. We anticipate that results from this work will provide a new perspective on the composition of oral biofilm communities and further insight into the role of interbacterial adhesion in dental plaque biofilm development.